DISCUSSION PAPER NO. 1613 Issues for Civil Service Pension Reform in Sub-Saharan Africa Anita M. Schwarz and Miglena Abels November 2016 Issues for Civil Service Pension Reform in Sub-Saharan Africa November 2016 Anita M. Schwarz and Miglena Abels Abstract The paper summarizes the main factors behind the projected increase civil service pension costs in Sub-Saharan Africa (SSA). It discusses the benefits and potential downsides to unifying civil service and national pension systems, drawing on regional and international best practices and experience. The paper pays special consideration to the differences between civil and national pension reform, emphasizing the unique challenges in civil service pension reform posed by the fact that the government is the employer, the administrator of the pension fund, and the guarantor of last resort of the pension system. Findings in the report strongly suggest that civil service pension reform needs to be on the agenda in SSA countries, as its costs are beginning to crowd out other budget expenditures. Among other conclusions and recommendations, the report also urges practitioners to focus on the overall impact on government finances and not on the finances of the pension fund when undertaking civil service pension reform separately from the national system. The paper is intended to serve as a resource in civil service pension reform efforts in the region. JEL Classification: H55, J26 Key Words: national and civil service pension systems in Africa, public pension system expenditure, design and performance indicators of civil service pension systems Acknowledgements This paper was prepared by a World Bank team led by Melis U. Guven (Sr. Social Protection Economist, Social Protection and Labor Global Practice) with financing from Rapid Social Response Multi-donor Trust Fund. 1 The paper was authored by Anita M. Schwarz (Lead Economist, World Bank) and Miglena Abels (Research Analyst, World Bank). The authors gratefully acknowledge the guidance and useful contributions of Melis U. Guven in the preparation of this paper. The peer reviewers were Fiona Stewart (Lead Financial Sector Specialist) and Mitchell Wiener Sr. Social Protection Specialist) and their comments as well as comments provided by Robert Palacios (Global Pensions Lead) greatly benefited the final document. The paper was prepared under the overall guidance of Dena Ringold (Practice Manager, Social Protection and Labor for East and Southern Africa) and Stefano Paternostro (Practice Manager, Social Protection and Labor, West Africa). Khurshid Banu Noorwalla (Program Assistant) provided administrative support. The findings, interpretations, and conclusions expressed in this document are those of the authors and do not necessarily reflect the views of the Executive Directors of the World Bank, the governments their represent, or the counterparts consulted during the preparation of the document. 1 The Rapid Social Response (RSR) program includes two trust fund programs, the RSR Multi-Donor Trust Fund (RSR-MDTF with initial contributions from the Russian Federation and Norway) and the RSR Catalyst Trust Fund (RSRC, with the United Kingdom as the sole contributor). During 2012-2013, three new donors joined the RSR- MDTF; namely Sweden, Australia, and the United Kingdom. i Table of Contents Introduction .............................................................................................................................. 1 1. Projected Explosion of Civil Service Pension Costs.......................................................... 2 2. Integration vs. Separate Civil Service Pension Systems ................................................ 15 3. Mechanics of Civil Service Pension Reform ................................................................... 21 4. Conclusions...................................................................................................................... 29 References .............................................................................................................................. 31 Annex ...................................................................................................................................... 33 A. Civil Service Pension Scheme Contribution Rates (% Wage) ................................... 33 B. Financing Mechanisms of Civil Service Pension Schemes........................................ 33 C. Evolution of Civil Service Pension Expenditures in Kenya ....................................... 34 Boxes Box 1: The Nigerian Pension Reform Experience ................................................................... 20 Tables Table 1: Wage Base for Pension Benefits in Select Sub-Saharan African Countries ................ 8 Table 2: Post-retirement Pension Indexation Measures ........................................................ 10 Table 3: Structure and Design of Civil Service Pension Schemes in Africa ............................. 16 Figures Figure 1: Population over Age 65 in Sub-Saharan Countries .................................................... 2 Figure 2: Pension Spending as a Percentage of Health Spending ............................................ 3 Figure 3: Pension Spending as a Percentage of Education Spending ....................................... 4 Figure 4: Pension Spending as a Share of Government Revenues ........................................... 4 Figure 5: Civil Service Pension Beneficiaries as a Percentage of Population above Statutory Retirement Age ......................................................................................................................... 5 Figure 6: Hypothetical Pension Spending as a Percentage of GDP if All Elderly Receive Civil Service Pension Benefits ........................................................................................................... 6 Figure 7: Accrual Rates in Sub-Saharan Africa and the OECD .................................................. 7 Figure 8: Expected Duration of Retirement—Life Expectancy at Age 60 ............................... 13 Figure 9: Demographic Structure of Public Service Pension Fund and National Social Security Fund in Tanzania ..................................................................................................................... 14 Figure 10: Administrative Costs as a Share of Contribution Revenue .................................... 24 ii Introduction Despite the youthful demographic of most Sub-Saharan African countries, many of their civil service pension schemes are becoming expensive. The majority of the countries in Sub- Saharan Africa for which there is data have separate pension schemes for public sector workers. Seven countries have top-up systems, with civil servants eligible for national pensions but also for an additional pension based on their civil service employment. Most of these countries require contributions from civil servants.2 Only five do not. Although the lifecycle of pension funds is still in its early stages in many of these countries, costs in some countries are already high. In Tanzania and Madagascar, civil service pension spending has reached 1.63 percent and 1.4 percent of GDP, respectively. It is lower in other countries, like Uganda, where it is 0.4 percent of GDP. To look at costs from another angle, in some places civil service pension spending is on the way to consuming as much of the budget as critical national priorities. For example, pension payouts are 33 percent of education spending and 25 percent of health spending in Mauritius. Civil service pension spending in Africa may not appear high relative to other countries  but it is extraordinarily high when we consider the number of people receiving pensions. In Uganda, for example, 0.4 percent of GDP is spent on civil service pensions, but this spending covers only 2.5 percent of the elderly population and an even smaller share of the overall population. If all of Uganda’s elderly received such pensions, payouts would be more than 16 percent of GDP, significantly higher than the 9.5 percent of GDP spent by high income countries with much older pension schemes. The generosity of the civil servant schemes is particularly important since civil service schemes often set the expectations for private pensions. Civil service pension reform is complicated by the fact that the government takes multiple roles: employer, fund administrator, and financier of last resort. As a result of this 2 See Annex for more information on funding mechanisms and contribution rates of civil service pension schemes in Sub-Saharan Africa. 1 multiplicity of roles, some standard reform prescriptions may not have the desired impact in the context of civil servant schemes, and different policy prescriptions may be required. The remainder of this note is organized in four sections: Section II, The Projected Explosion of Civil Service Pension Costs; Section III, Pros and Cons of Unifying Civil Service and National Pension Systems; Section IV, Special Considerations When Reforming Civil Service Pension Systems; and Section V, Conclusions. 1. Projected Explosion of Civil Service Pension Costs Sub-Saharan Africa is generally thought of as a young region, but the percentage of the population over age 65 is expected to double in most countries and rise by an even greater percentage in a few of them. Figure 1 shows the percentage of population over the age of 65 in 2010, compared to what is projected for 2050. Given that the population is expected to double in this timeframe, this indicates that the number of elderly in Africa are expected to quadruple. Figure 1: Population over Age 65 in Sub-Saharan Countries (Percentage of Total) 30% 25% 20% 15% 10% 5% 0% Sao Tome and… Somalia Gambia Chad Guinea-Bissau Côte d'Ivoire Guinea Sudan Zambia Comoros Ethiopia Benin Eritrea Malawi Lesotho Rwanda Namibia Burkina Faso Niger Burundi Congo, Dem. Rep. Kenya Djibouti South Africa Sierra Leone Ghana Cabo Verde South Sudan Seychelles CAR Mozambique Tanzania Togo Swaziland Liberia Gabon Uganda Madagascar Mauritius Mali Senegal Botswana Nigeria Zimbabwe Angola Cameroon 2010 2050 Source: United Nations 2015. Civil service pension costs in Sub-Saharan Africa are high relative to other social expenditures and growing. Figure 2 and Figure 3 compare spending on civil service and 2 national pension schemes to health and education spending in 17 Sub-Saharan countries for which there is data. Overall, the amount spent on pensions represents an average of 18 percent of health spending and 28 percent of education spending. Of the overall pension spending, 70 percent goes to civil service pensions. Pension spending costs as a share of government revenues are generally in line with experience in other developing economies. However, spending levels are high considering the small share of elderly in populations across the region (Figure 4). Figure 2: Pension Spending as a Percentage of Health Spending 70% 60% 50% 40% 30% 20% 10% 0% national pension spending, % health spending civil service pension spending, % health spending Source: National and civil service pension spending−administrative data; Education and Health spending−World Development Indicators. Note: Data presented is for the latest year available. 3 Figure 3: Pension Spending as a Percentage of Education Spending 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% national pension spending, % education spending civil service pension spending, % education spending Source: National and civil service pension spending−administrative data; Education and Health spending−World Development Indicators. Note: Data presented is for the latest year available. Figure 4: Pension Spending as a Share of Government Revenues Pension Spending as a Share of Government Revenues vs. Population 65+ as a Share of Total Population 9% 8% Mauritius NPF 7% population 65+, % total Seychelles 6% 5% Ghana SSNIT Swaziland SPF South Africa GEPF 4% Ethiopia POESSA Botswana BPOPF 3% Liberia NPS Kenya NSSF 2% Togo Civil Service Uganda NSSF Madagascar CRCM Benin FNRB 1% Namibia GIPF 0% 0.0% 2.0% 4.0% 6.0% 8.0% 10.0% 12.0% 14.0% pension spending, % government revenues Source: Administrative data from respective countries; IMF 2015; and United Nations 2015. 4 Civil service pensions cover a fairly narrow portion of the elderly. Figure 5 shows the number of beneficiaries above the statutory retirement age of the country (commonly 60 for men and women) as a percentage of the population. The percentage of the elderly that actually receive pensions is probably even lower than the numbers shown because in most cases, the data covers pension beneficiaries irrespective of age. Many of these beneficiaries are younger than retirement age because of early retirement provisions, suggesting that even fewer people above retirement age are actually collecting pensions. Figure 5: Civil Service Pension Beneficiaries as a Percentage of Population above Statutory Retirement Age Percentage of Elderly Receiving Pensions 16.00% 14.00% 12.00% 10.00% 8.00% 6.00% 4.00% 2.00% 0.00% Source: Administrative data from the respective countries. If current civil service pension benefits were expanded to cover all elderly, African countries would spend more on pensions than high income countries with much larger elderly populations. Figure 6 illustrates what would happen if civil service pension benefits were expanded to cover all the elderly in Ghana, Tanzania, Uganda, and Zambia, and compares it to pension spending as a percentage of GDP in high income countries. High income countries are having difficulty financing pensions out of their much larger incomes and tax collections. It is hard to imagine how poorer countries could spend an even larger share of their incomes on pensions. While this is purely hypothetical, there are some lessons to be drawn. 5 First, civil service pension benefit levels are more generous in Sub-Saharan Africa than in high income countries. Generous civil service benefit levels also tend to push up benefit levels for the private sector, as the private sector tries to compete with the public sector for the most competent workers. While countries can afford the generous benefits now, they may become unaffordable in the future. It takes time to reduce benefit levels, since pensions are often considered acquired rights and are sometimes even constitutionally protected. Second, if the costs to cover all elderly are already high as a percentage of GDP, they will be unimaginably high when the number of elderly doubles or even quadruples. So again, it is worthwhile to think about reforming these civil service systems now when there is still time to do so. Figure 6: Hypothetical Pension Spending as a Percentage of GDP if All Elderly Receive Civil Service Pension Benefits 18 16 14 12 10 8 6 4 2 0 Ghana Tanzania Uganda Zambia High Income Source: Staff calculations. Civil service pensions are expensive because benefits are generous and pensions are paid out over a long period, in part due to early retirement. Generous benefits can arise from various parameters of a pension scheme: (i) the accrual rate, the rate at which the benefit entitlements build up for each year of service; (ii) the averaging period for wages included in the pensionable wage; (iii) indexation of the pension post-retirement; and (iv) commutation options. 6 Pensions in Sub-Saharan African accrue at an average rate of 2.2 percent of salary per year of service, compared with an average rate of 1.2 percent in high income countries. The accrual rates in high income countries, which are considered more or less fiscally sustainable and international best practice, are nearly half of average accrual rates in Sub-Saharan African countries. The African accrual rates translate into 66 percent of pensionable wage after 30 years of service. Kenya’s and Togo’s civil service pension schemes have accrual rates of 2.5 percent, promising to deliver replacement rates equal to 75 percent of previous earnings after only 30 years of service. The accrual rate is even higher in Mozambique, where public sector workers can expect their pensions to equal 86 percent of their wages. Furthermore, the accrual rates in civil service schemes in Tanzania, Mozambique, and Togo are higher than the accrual rates in the national pension systems, creating a disparity between civil servants and private sector workers. The actual pension benefits received by a high income retiree may not be significantly higher than for an African civil servant retiree, because the high income retiree is expected to work 40 or more years to receive a full pension, rather than the 30 or, in most cases fewer years, typical in Sub-Saharan Africa. Figure 7: Accrual Rates in Sub-Saharan Africa and the OECD 3.5% 3.0% 2.5% 2.0% 1.5% 1.0% 0.5% 0.0% United States Iceland Turkey Zambia Madagascar Kenya Mozambique Slovak Republic Canada Norway Austria Netherlands Poland Korea Slovenia Germany Spain Tanzania Namibia Uganda Cabo Verde Congo, Dem. Rep. Estonia Japan Sweden Belgium Italy Luxembourg Togo Cameroon Swaziland SSA - Civil Service Pension Schemes OECD - National Pension Systems Source: Staff notes; OECD 2015. 7 African civil service pensions are often based on last salary or on the last few years of salary, in contrast to international best practice that bases pensions on average lifetime wages. Another component of the benefit formula that influences generosity is the number of years in the averaging period on which pension benefits are calculated. Many civil service pension schemes in Sub-Saharan Africa base pensions on either last wage (Cabo Verde, Uganda, Madagascar), the average of the best five years of wages (Sierra Leone), or the average of the best three years of wages (Ghana) (Table 1). The logic behind basing pensions on lifetime wages is based partly on fairness and partly on meeting a worker’s expectations. If a worker has paid contributions on earned wages over a lifetime, then aligning benefits with contributions would imply that pensions should be based on the average lifetime wage, since the average contribution was based on the average lifetime wage. Presumably, one’s consumption patterns are based on lifetime average wages and not just on final salary. These historical wages are typically adjusted upward for inflation and typically also include upward adjustments for growth in average economy-wide wages. Including average wage growth in the calculation of past wage history is seen as providing an “interest rate” on past contributions, much like an interest rate is paid on a bank account. Table 1: Wage Base for Pension Benefits in Select Sub-Saharan African Countries Pension Scheme Country Wage Base national Benin final 5 both national and civil service Ethiopia final 3 civil service Swaziland final salary integrated Seychelles last 5 civil service Mauritius last 5 integrated Cabo Verde final salary civil service Madagascar final salary national Sierra Leone best 5 avg. national Togo final salary civil service Uganda final salary national Mozambique final salary national Ghana best 3 avg. civil service Tanzania final salary national Tanzania best 5 avg. civil service Namibia final salary civil service Zambia final salary Source: Administrative data from respective countries. 8 Civil servant pensions based on final salary can be costly since civil services wages increase steadily each year and substantial promotions may come at the end of a career. Typically, civil servants see steady increases in their salaries, as salaries are often based on seniority. Private sector salaries, by contrast, often peak earlier in the career, around the late 40s or early 50s, and then either hold steady or decline slightly. As a result, the overpayment in benefits relative to average contributions noted above tends to be much higher for civil servants. Furthermore, in many cases, civil servants receive promotions and large salary increases just before retirement, which bumps up their pensionable salaries and results in huge costs to the pension system. Post-retirement indexation of benefits in Sub-Saharan Africa typically increases their generosity. International best practice suggests indexation post-retirement should be by inflation only, with the logic that an individual’s purchasing power should be maintained from the first day of retirement throughout retirement, however long it lasts. However, the vast majority of countries in Sub-Saharan Africa adjust civil service pensions to reflect civil service wage growth or on an ad-hoc basis to approximate civil service wage growth, which is typically higher than inflation (Table 2). Some countries go further and tie post-retirement benefits to the wages of individuals who hold the same position as the worker last held. Thus, if an individual retired as Permanent Secretary, his pension during retirement will increase based on the increases in wages experienced by current Permanent Secretaries. Since wages at the upper echelons of the civil service typically rise faster than for the average civil servant, this type of indexation can further increase costs. 9 Table 2: Post-retirement Pension Indexation Measures Country National Civil Service Cabo Verde ad-hoc ad-hoc Kenya inflation Namibia higher than inflation Seychelles inflation Tanzania ad-hoc ad-hoc Uganda wages Zambia wages ad-hoc Ghana wages Benin wage Ethiopia none none Madagascar ad-hoc Sierra Leone wages Togo no indexation wages South Africa at least 75% inflation Sao Tome inflation Source: Administrative data from respective countries. Commutation of pensions further increases costs, at the same time increasing the likelihood that civil service pensioners will face poverty at the end of their life spans. Commutation, a common feature of colonial expatriate benefits, allows individuals to take a portion of their pension as a lump-sum at the beginning of their retirement period. For expatriate civil servants, this was often an important feature which allowed them to purchase a home in their homeland after having served abroad during their working years. Technically, individuals who choose commutation should receive the present value of the portion of their pension that is commuted and is based on the average life expectancy at the time of commutation. Since the commutations are typically spent right away, the pension available for the retiree is reduced, and particularly, if the retiree lives a long time, retirees may find themselves facing poverty. Generous commutation factors further increase the generosity of the pension. In practice, the conversion from a pension to a lump-sum is done using a commutation factor, which is often fixed in legislation. The life expectancy of civil servants as a distinct group is typically unknown. What is known across a broad range of countries is that high income individuals 10 tend to live much longer than lower income individuals, and African civil servants are typically high income or upper middle income individuals in their respective countries. A preliminary version of a Social Protection Policy Note, which uses data from formal sector workers in Ghana, suggests that mortality rates during retirement may be more than 25 percent lower than UN national averages for males and more than 60 percent lower for females in the formal sector.3 As a result, even actuarially fair commutation factors based on national life expectancies are likely to produce very generous commutations for civil servants. The factors are also often written in law and do not adjust as life expectancy or other pension system parameters change. Furthermore, some countries have a restoration feature. They consider that an individual has “bought,” for example, 12 years of reduced pension payments in advance by receiving a lump sum. Should the individual live longer than 12 years, the full pension is restored. What this restoration ignores is that if the life expectancy were actuarially fair, for every individual living beyond 12 years, there would be a person who lives fewer than 12 years, and the commutation balances the overpayment to the shorter-lived person with the underpayment to the longer-lived person. Restoration removes that balancing mechanism, making the pension system more expensive. Pension benefits in Sub-Saharan Africa are paid out for more years than average. It is a well- documented fact that life expectancy in Africa is lower than in other regions of the world. Sub-Saharan African countries have an average life expectancy at birth of 54 years, while high income countries have a life expectancy at birth of 79.4 years, a difference of slightly more than 25 years. However, much of the mortality occurs at infancy and childhood: those who reach retirement age in Sub-Saharan Africa still have a lower life expectancy than those in high income countries, but the gap closes to less than seven years when looking at life expectancy at age 60. While the normal retirement age in the civil service schemes in Sub- Saharan Africa tends to be 60 for both men and women, a number of countries allow early retirement with no or limited reductions in benefits. In Rwanda, people are allowed to retire as early as 40, while in Botswana, Uganda, and the Gambia, early retirement is allowed as 3 Majoka and Palacios 2014. 11 early as 45. Kenya allows early retirement at 50 and Tanzania at 55. High income countries allow full retirement only at 65 in some cases; in others 67 or 68. This creates a 28-year gap between the earliest civil service retirement age in an African country and the oldest minimum retirement age in a high income country. The difference in life expectancies cannot justify such a huge gap. Even at age 60, the duration of retirement would be well above the international best practice of 15 years in all Sub-Saharan African countries but Sierra Leone and the Gambia, as shown in Figure 8. It should also be noted that the life expectancies shown in Figure 8 are for the national populations, which as discussed above underestimate the life expectancies of civil servants. As civil service pension systems in Sub-Saharan Africa mature, costs are expected to increase even further. When pension systems first begin, they have relatively large numbers of contributors, but few beneficiaries. Typically, in order to qualify for a benefit, an individual needs to have contributed or belonged to the system for a minimum number of years. Once that minimum number of years has passed, cohorts of individuals begin to retire. However, each of these individuals will have barely met the minimum criterion to qualify for a pension and would be unlikely to qualify for a full pension. Only 30 or 40 years after inception of a pension system does the first cohort of employees with a full career of contributions begin to retire. Even when this cohort retires, the bulk of pensioners receive smaller pensions since they only had partial contribution careers and thus partial pensions. It is only when all pensioners are likely to be full contributors that the pension system reaches maturation and the full costs of the pension system become evident. The time for this maturation depends on how long it takes for a new entrant to the civil service to reach retirement and the life expectancy after retirement. System maturation thus is expected to take 50–70 years depending on the country. Given that many of the African countries obtained their independence after World War II and began to set up their civil service systems after independence, many of the countries are reaching system maturation only now. 12 Figure 8: Expected Duration of Retirement—Life Expectancy at Age 60 Duration of Retirement, Male Duration of Retirement, Female 85 90 80 85 75 80 70 17 17 75 22 16 1515 16 18 16 15 16 17 13 16 18 1717 22 65 11 70 17 17 16 18 19 18 16 17 21 14 19 18 18 1616 65 17 17 17 11 60 60 55 55 50 50 45 45 40 40 Zambia Benin Sierra Leone Rwanda Kenya Togo Liberia Cabo Verde Ghana Tanzania Seychelles Mozambique Uganda Madagascar Guinea Bissau Mauritius The Gambia Nigeria Botswana Zambia Benin Rwanda Sierra Leone Kenya Togo Cabo Verde Liberia Mozambique Ghana Tanzania Uganda Seychelles Madagascar Guinea Bissau Mauritius The Gambia Nigeria Botswana retirement age life expectancy at age 60 retirement age life expectancy at age 60 Source: Retirement age—administrative data; Life expectancy at age 60—United Nations, latest update of December 2012. Prior to system maturation, even if the parameters of a pension system are actuarially unbalanced, the system will show relatively low expenditures and relatively high revenues. Prior to maturation, a pension system has many contributors and few beneficiaries, with the beneficiaries it does have receiving lower benefits due to making only few years’ worth of contributions. After maturation, both benefit levels and the number of beneficiaries increase but revenue levels remain similar to what they were before. This switch can make the actuarial imbalance in the pension system more obvious. The demographic structure of many African civil service pension systems is also much older than the general population, suggesting that contribution-based civil servant systems will begin to show deficits long before the comparable national systems. It seems odd to speak of the African civil services as suffering from aging, but in fact, the civil services expanded in the 1960s and 1970s. While increases in some civil service positions, such as teachers and police, are often tied to population growth, in other cases the size of the civil service is affected by other factors, such as budgetary expansions and slowdowns. After an initial ramping up of the civil service, most countries have slowed growth due to budgetary 13 pressures. With fewer young, new entrants, and aging of the existing civil servants who live longer than the general population, the age structure of participants often looks more like that of an aging country than of a young country, with all the fiscal problems that older countries face in financing their pensions. Contribution revenues are declining while pension expenditures continue to rise. For example, the demographic structure of the Public Service Pension Fund in Tanzania is significantly older than that of the National Social Security Fund (Figure 9). The system dependency ratethe ratio of old age beneficiaries to effective system contributorsis more than three times as high in the civil service scheme than it is in the national pension scheme. Figure 9: Demographic Structure of Public Service Pension Fund and National Social Security Fund in Tanzania 30% 25% SYSTEM DEPENDENCY RATE 20% 15% PSPF NSSF 10% 5% 0% 2016 2019 2022 2025 2028 2031 2034 2037 2040 2043 2046 2049 2052 2055 2058 2061 2064 2067 2070 2073 2076 2079 Source: Staff calculation. For all of the reasons given above, civil service pension schemes in Sub-Saharan Africa are already costly given the limited number of people they cover, and their costs are expected to accelerate over time. The costliness of the civil service can largely be attributed to the following features: (i) generosity in benefit levels per year of service; (ii) benefits based on final salary or salary in a very limited number of years; (iii) generous commutation rules; (iv) long duration of benefits, both because of long life expectancies and prevalence of early retirement; (v) system maturation; and (vi) “aging” of the civil service in contributory systems. 14 2. Integration vs. Separate Civil Service Pension Systems Most countries in Sub-Saharan Africa have separate civil service systems. Of the 44 countries reviewed, 33 have separate schemes for public sector workers. In seven countries (Botswana, Lesotho, Mauritius, Namibia, South Africa, Liberia, and Swaziland) public sector workers are typically covered by some type of social pension, in some cases a means-tested pension, in addition to their civil service pension. Eleven countries (Cape Verde, Central African Republic, Chad, Ethiopia, Ghana, Nigeria, Rwanda, Sao Tome e Principe, Seychelles, Sierra Leone, and Zambia) have an integrated pension system. In the case of Cape Verde, integration of the national and civil service pension schemes has been done quite recently. Eritrea and South Sudan have not yet established private sector pension schemes. International experience reveals that fully or partially integrated pension systems are strongly preferable to a separate public sector pension system. Only nine OECD countries have entirely separate institutions and benefits for civil servants. Seven countries (Chile, Czech Republic, Estonia, Hungary, Mexico, Poland, and Slovak Republic) have completely integrated pension systems. In five countries (Denmark, the Netherlands, Finland, Iceland, and Israel), there are institutionally separate pension systems that provide very similar benefits. Another set of countries (Australia, Canada, Ireland, Italy, Japan, New Zealand, Norway, Slovenia, Spain, Sweden, Switzerland, and the United States) have fully integrated systems with top-up arrangements for civil servants. The United Kingdom presents a rather complicated example of a partially integrated system with a top-up scheme. It is important to emphasize that while many developed countries still often maintain some separate systems, there has been a notable movement toward more integration. The optimal design for a civil service pension system may be to integrate it with the national system. Integration typically allows better labor market mobility and tends to avoid some of the extreme generosities in civil service systems. 15 Table 3: Structure and Design of Civil Service Pension Schemes in Africa Structure of Civil Service Design of Civil Service Scheme No. Country Pension Scheme (if separate, otherwise national) 1 Botswana Separate FDC 2 Cabo Verde Integrated PAYG DB 3 Guinea Bissau Separate PAYG DB 4 Kenya Separate DC 5 Lesotho Separate FDC 6 Namibia Separate FDB 7 Rwanda Integrated PAYG DB 8 Seychelles Integrated PAYG DB 9 Tanzania Separate PAYG DB 10 Uganda Separate PAYG DB 11 Zambia Integrated PAYG DB 12 Zimbabwe Separate PAYG DB 13 Ghana Integrated Hybrid 14 Mauritius Separate PAYG DB 15 Angola Separate PAYG DB 16 Benin Separate PAYG DB 17 Cote d'Ivoire Separate PAYG DB 18 Ethiopia Integrated PAYG DB 19 Gambia, The Separate PAYG DB 20 Liberia Separate PAYG DB 21 Madagascar Separate PAYG DB 22 Mozambique Separate PAYG DB 23 Niger Separate PAYG DB 24 Nigeria Integrated FDC 25 Sao Tome and Principe Integrated PAYG DB 26 Senegal Separate PAYG DB 27 Sierra Leone Integrated PAYG DB 28 South Africa Separate FDB 29 Swaziland Separate FDB 30 Togo Separate PAYG DB 31 Burkina Faso Separate PAYG DB 32 Burundi Separate PAYG DB 33 Cameroon Separate PAYG DB 34 Central African Republic Integrated PAYG DB 35 Chad Integrated PAYG DB 36 Congo, Dem. Rep. Separate PAYG DB 37 Congo, Rep. Separate PAYG DB 38 Malawi Separate PAYG DB 39 Mali Separate PAYG DB 40 Mauritania Separate PAYG DB 41 Guinea Separate PAYG DB 42 South Sudan Separate PAYG DB 43 Sudan Separate PAYG DB 44 Eritrea Separate Source: Administrative data from respective countries. 16 Three advantages of a separate, generous civil service pension system are often cited: (i) a generous pension package back loading the overall compensation package induces people to remain in the civil service and discourages corrupt practices; (ii) a mandatory retirement age for civil service workers can accompany the pension plan; and (iii) a generous pension package attracts individuals to civil service. The rationale for the first is that if the compensation package is back loaded, and a civil servant is found guilty of corruption, he or she has more to lose than if the compensation package were structured differently. On the second point, civil servants are more likely to remain with the same employer throughout their careers: salary increases are almost always tied to tenure in the job, since productivity in the public sector is hard to measure. Civil servants also are more likely to want to continue working because of non-salary benefits they would lose upon retirement. Mandatory retirement ages, which are rarely appropriate in private sector pension schemes, might have a place in public employment pension schemes. In countries without mandatory retirement, there are examples of 80-year-old civil servants coming to work and sleeping most of the day just to collect a paycheck. The third reason for separate systems in high income countries with vibrant private sectors is that the public sector often pays lower salaries than the private sector. It uses generous benefits as a mechanism to continue to attract high quality workers, mostly because it is easier for politicians to promise higher future benefits than to pay higher salaries today. Lower-income countries adopted the pension and compensation structures of higher income countries, but in many of these countries, the public sector is the most lucrative form of employment. As a result, these countries provide an unnecessary incentive, since civil service employment typically pays both higher wages and higher benefits than the private sector. The advantages of a unified national system typically far outweigh the advantages of a separate system in most countries, by removing impediments to labor mobility. Many types of work are not specific to the public sector; clerical work and administrative tasks are often similar in both the private and public sectors. 4 Setting up a separate pension system for public 4 See Sluchynskyy 2015 for a detailed analysis of pension system administrative costs. 17 sector workers discourages workers from pursuing opportunities across sectors because eligibility for benefits are usually dependent on working or contributing for a minimum number of years to the pension system, with benefits dependent on how long a worker contributes. Careers spent in different pension systems invariably end up with workers losing some of the benefit they would otherwise have been entitled to. In some larger countries, where municipal or state workers have their own pension systems separate from federal workers, the separate pension systems even discourage teachers and health workers from moving from one state or city to another. These issues are even more important for developing economies than for developed ones as the formal sector is still evolving, with new job opportunities still developing and with new workers moving to urban areas. A unified national system allows workers to move to the best job opportunities without fears of losing pension rights. 5 It also gives the public sector more flexibility to downsize or right size its workforce, given that workers have more opportunities for other types of employment. Various limitations might prevent public and private systems from being unified in the short and medium term: (i) There may not be a viable private sector pension system for public sector employees to join; (ii) the private sector system might also be overly generous, or there might be a tendency to apply the more generous public sector provisions to the private sector as well; (iii) the transition to a unified system may require unavailable fiscal resources; (iv) policymakers may be reluctant to combine systems running financial surpluses with those running deficits; 6 and (v) it may be difficult to reconcile any redundant administrative capacity. In most countries, civil servants are the first group to be covered by a pension system. This is partly because civil service careers tend to be stable and continuous, making benefit administration easier than in the private sector, where compiling contributions from various employers and tracking employees as they change jobs can be more challenging. As a result, there may be no comprehensive private pension system. Another problem is that the private sector system might not offer comparable benefits, 5 See Palacios and Whitehouse 2006. 6 See Palacios forthcoming. 18 resulting in pressure to bring it up to the level of the civil pension scheme. Alternatively, the private scheme might also offer generous benefits, which could mean increased future liabilities if the civil sector is integrated with it. Transitioning from one system to another might result in unaffordable fiscal costs, particularly if the private sector system has lower contribution rates or is set up as a savings system. If the private sector system is newer with fewer beneficiaries, it may be reluctant to take on a public sector system that is already running deficits; or a public system with surpluses might be reluctant to join a private sector system with deficits. The administrators of each of several pension funds may also have concerns about whether unification will result in redundancy of jobs. As a result, the policymaker is often tasked with reforming the civil service pension system separately from the national system. In many cases, policymakers will undertake reforms in the civil service pension system apart from the national system. However, it is always worthwhile looking at what exists for private sector workers and trying to see what can be harmonized between the two. This may improve some of the labor mobility issues and help set the stage for further integration in the future. As swelling pension costs increase pressures on national fiscal accounts (particularly in civil service pension schemes), some Sub-Saharan countries have undertaken reforms aimed at containing costs and improving the overall efficiency of the social insurance sector. In some cases, the reforms have involved a conversion from a PAYG DB scheme to a funded DC scheme (Lesotho). Other countries have enacted pension reforms aimed at integrating national and civil service schemes (as Cabo Verde has done recently) without changing the financing mechanism or design of the pension system. Nigeria accomplished one of the most ambitious and radical pension reforms in the region. The Pension Reform Act of 2004 did away with the PAYG DB system for public sector workers and established a fully-funded defined contribution (FDC) integrated pension system for public and private sector workers alike (Box 1). Namibia (Government Institutions Pension Fund—GIPF) and South Africa (Government Employees Pension Fund—GEPF) provide examples of well-managed funded DB schemes. 19 Box 1: The Nigerian Pension Reform Experience Prior to 2004 reform, the Nigerian pension sector was characterized by extensive fragmentation and heterogeneity of eligibility conditions, financing mechanisms and benefit levels between the various public and private sector pension programs. In addition, the pension system had accumulated large-scale pension arrears in the public sector and had consistently failed to pay out promised pensions. The first national pension scheme in Nigeria was established after the country gained independence in the early 1960s. It was developed out of the provident fund scheme that had catered to colonial civil service: a severance payment scheme that paid a lump sum on retirement. In 1994, the National Social Insurance Trust Fund (NASTF) was created—the first in-country pension scheme that paid out annuities. NASTIF extended pension coverage only to private sector workers. On the other hand, the NASTIF scheme was a PAYG scheme financed from employee and employer contributions. Numerous parallel civil service schemes also served different branches of the public sector. The pension schemes for federal, state and local civil servants were non-contributory and unfunded. The differences between public and private sector pension schemes extended far beyond the financing mechanisms. Public sector pensions could be obtained earlier than the standard retirement age if a certain number of years of service were met; they were also far more generous than private sector pensions. Not surprisingly, pension spending for retired public sector workers greatly exceeded pension spending for private sector retirees. As a result of the different eligibility conditions and generosity levels between private and public sector pension system, among numerous other problems, Nigeria faced an inequitable pension sector and rapidly rising civil service pension costs. The Pension Reform Act of 2004 ushered in a mandatory funded contributory pension scheme (FDC) for employees in the public and private sectors, a reform largely regarded as the most radical reform initiative in the region to date. Key elements of the reform included: • An Individual Retirement Savings Account (RSA), which every employee was to open and maintain in his/her name with any pension fund administrator of his choice (total contribution rate is 18 percent, 10 percent for employers and 8 percent for employees); • Licensed Pension Fund Administrators (PFAs) were to open retirement savings accounts for employees, invest and manage the funds in fixed income securities and other instruments as may be determined by the Regulatory Agency, the National Pension Commission (NPC); • Pension Assets Custodians (PACs) were licensed to warehouse pension fund assets; • A Retirement Benefits Bond Redemption Fund, maintained by the Central Bank of Nigeria, covers the gap created by the switch from the DB scheme to a DC funded system. The Federal government credited the Fund with an amount equal to 5 percent of the total monthly wage bill of federal workers. The Fund also settled the pension debt accrued under the old scheme; • The National Pension Commission (NPC) was introduced as the regulatory body that oversees the new contributory scheme. (Continued) 20 Box 1: The Nigerian Pension Reform Experience (Continued) The pension industry witnessed 1.76 percent growth in the scheme membership during the second quarter of 2016, increasing from 6,581,031 contributors at the end of the preceding quarter to 6,696,793. The expansion in industry membership was solely driven by the RSA Scheme. Membership in the Closed Pension Fund Administration Scheme (CPFA) experienced negative growth, while membership of the Approved Existing Scheme (AES) remained unchanged. The private sector continued to lead in RSA membership, with 52.13 percent (3,457,029) of total RSA registrations as at the reporting period. The sector also witnessed a higher growth of 2.17 percent (73,573) than the public sector. This can be attributed to an increased level of compliance by the private sector as a result of the various steps taken by the Commission to improve compliance and coverage, as well as marketing strategies of the PFAs. Although this appears to be a positive development in Nigeria’s pension sector, detailed evaluation of the performance of pension funds is not possible because of a lack of information on investment returns and generally limited public availability of data. In addition, the pension sector in Nigeria continues to operate in a largely informal 7 labor market environment, which poses significant implications for extending coverage to workers who fall outside the formal labor market. Nigeria also continues to face challenges related to governance and transparency, which can negatively affect the performance of the pension system. The 2014 Pension Reform Act (PRA), signed into law on July 1, replaces the social security pension system established in 2004 with a new pension system modeled on the current one. Some of the key changes to the current system include closure of stand-alone pension plans (provided in lieu of participation in the social security plan) to new members and an increase in contribution rates (previously contribution rates were set at 15 percent, evenly split between employer and employee). In addition, small employers (fewer than three employees) and the self-employed are now permitted to participate in the pension system on a voluntary basis. Source: Watson 2014; Nigeria National Pension Commission 2016. 3. Mechanics of Civil Service Pension Reform Since civil service pension reform is often undertaken as a self-standing reform, it is important to understand the similarities and differences between it and national system reform. While the mechanics of civil service pension reform may mimic the mechanics of national pension reform, key differences arise from the fact that the government is the employer, the administrator of the pension fund, and the guarantor of last resort of the pension 7 For more the labor market situation in Africa, see Filmer & Fox 2014. 21 system. The many different roles that the government plays in a civil service pension system lead to conflicting objectives and can sometimes lead to misguided reform attempts. Countries are often advised to move away from simple civil service pension systems where the government pays pensioners directly from the budget, and instead to set up a separate pension fund to which the workers and government as employer make contributions. National systems typically collect contributions from employers and their employees and hold them in a pension fund, paying benefits from this pension fund as the employee reaches retirement age. The pension fund is typically administered by government employees. Should the fund not have sufficient reserves to pay pensions, the government is legally obliged to pay benefits from general government revenues. In theory, the advantages are that the employee has a third party, the pension fund, making sure that the employer pays contributions on time, a third party managing the reserves and holding them so even if the employer goes bankrupt the funds will be there, and a fourth party, namely the government, guaranteeing that the legally defined benefits will be paid. And given that the government has an interest in limiting the amount of deficits it has to pay, it has some interest in making sure that the employers pay the contributions and the pension fund manages the reserves judicially. In practice, not all pension funds actually realize these advantages. When the government is also the employer, a separate pension fund may provide little additional protection for the worker. If the government is liable for pension payments when a worker retires, it really does not matter to the worker whether the government as employer makes contributions on a regular basis. If the government finds itself cash-strapped, it may be sensible to make only the contributions necessary to pay today’s pensions while recognizing that there will be additional payments necessary in the future. It makes very little sense for the government to borrow in the market in order to make contributions to the pension fund that will accumulate there as reserves. In addition, pension fund reserves might have become an additional asset of the government under the simple system. Whether the government accumulates reserves today or “invests” the reserves in politically important priorities becomes a question of whether the government prefers to pay today or pay in the 22 future. As might be imagined, governments frequently make the choice of paying in the future by running arrears to the civil servant pension fund or borrowing from the fund or having the fund “invest” in the government’s political priorities. Political investment of reserves can also take place in national pension funds, but typically the government needs to be more accountable to the employees and employers who are providing the financing than when the government is the prime financier. Creating a pension fund results in a new interest group, the pension fund administration, whose interests may not align with those of the workers or of the government as employer or as payer of last resort. The pension fund administration has separate interests from the government. Pension fund administrators gain political and economic clout by accumulating reserves. These reserves can be invested in the economy, with the pension fund administrator frequently having a lot of leeway in where and how they are invested. Control of these off-budget surpluses gives the pension fund administration both economic and political clout. As a result, pension fund administrations often request that contribution rates be raised, even if they are to sit in a surplus fund, or that deficit-running pension funds be closed, with the deficits being passed to the government, and a new pension fund opened that can accumulate reserves. The finance and labor ministries, which often are heavily involved in supervising pension funds in older, deficit-laden countries, may be less diligent in supervising pension funds that are running surpluses, as they are not perceived as causing problems for the Treasury. Furthermore, pension fund administrations sometimes spend heavily on administrative costs by investing in buildings and cars as well as perks for their own staff, none of which are in the best interests of the workers or the government. 23 Figure 10: Administrative Costs as a Share of Contribution Revenue 8 0.6 0.5 0.4 0.3 0.2 0.1 0 National Civil Service Source: Administrative data for each respective country and pension scheme; Sluchynskyy 2015. The only benefit of maintaining an actual pension fund for civil servants is the increased cost transparency of hiring additional civil servants. If the government as employer needs to make contributions to a pension fund, hiring an additional civil servant will clearly cost more than if the initial cost is only salary with the promise of a pension sometime in the future. The more actuarially balanced the pension system is, with expected contribution revenue equal to expected pension expenditures, the more transparency is added to civil service hiring decisions. Actuarially balanced schemes can also encourage policymakers to think twice before randomly increasing benefits, as additional costs will be borne by the budget immediately. However, most pension schemes are not actuarially balanced, reducing this advantage of maintaining a separate pension fund. Additional costs will also be borne immediately by the budget in the case of funded DB schemes, such as those in South Africa and Namibia. This is also the case in the context of fully funded DC schemes (Botswana and Nigeria). 8 See Sluchynskyy 2015 for more on administrative costs. 24 Even simple reforms such as raising contribution rates have different implications for national systems and civil service systems. In a national system, raising contribution rates typically brings in additional revenue as both employees and employers contribute higher percentages of salary to the pension system. Assuming that benefits are kept the same, raising contribution revenue typically reduces the deficits that the government needs to finance. If contribution rates are already quite high, raising rates may induce workers and employers to evade the system, resulting in little additional revenue; but typically it either increases surpluses or reduces deficits. In a civil service system, by contrast, raising employer contribution rates essentially raises government costs. Whether the government faces additional deficits or higher contribution costs is immaterial from its perspective. Even the ability to raise revenue by raising employee contribution rates is often compromised by the strength of civil service unions that object to a decline in net salaries. As a result, governments often end up raising wages of civil servants to compensate for the increased contributions, which completely counterbalances the positive impact of higher contributions on the overall government balance. So while the pension fund balance improves, the overall government fiscal picture might not. This is one of the many cases where the interests of the pension fund manager in maintaining surpluses might diverge from the interests of the government, which is interested in overall fiscal sustainability. In Uganda, for example, the civil service pension reform proposal included a one-time compensational pay raise in the form of a one-time 5 percent wage growth adjustment on top of the normal pay raise to compensate workers for a higher contribution rate. On the other hand, if the increased costs can be imposed primarily on employees without an accompanying wage increase, there might be some improvement in the fiscal picture because, unlike participants in national systems, civil servants are less able to evade contributions. In some cases, governments begin to hire workers as contractors rather than civil servants to avoid the higher cost of the civil servant employee. Raising the contribution rate for a system that is in surplus can do more harm than good. A rising contribution rate makes even less sense in a system that is accumulating reserves, because it can result in the government borrowing funds to finance the increased 25 contributions and increased wages, simply to generate a larger surplus in the pension fund. To be sure, while additional government costs contribute to a tighter fiscal constraint today, they potentially relieve future costs. But for those future savings to materialize, the savings need to be invested in a way that generates fairly high returns. Otherwise the government faces higher costs today and higher costs tomorrow. This is even truer if the government has to borrow the fiscal resources to pay for the increased contributions today. If the borrowed funds are just going to accumulate in a reserve fund, the fund has to earn a higher rate of return than the cost to the government of borrowing. The benefit of the higher contribution rates accumulating in a fund is that each generation is ostensibly paying for its own pensions. But if the reserves do not generate reasonably high returns, contributions for future workers or their taxes will have to increase to pay future pensions anyway. Increasing the retirement age, which benefits national pension systems, does not necessarily save money in civil service pension systems. In a private sector system, raising the retirement age increases the number of contributors, if the contributors choose to continue working, and thus increases the amount of revenue going into the pension system. At the very least, it reduces the number of beneficiaries, decreasing expenditures, thus bringing the pension system closer to balance. In a civil service system, the employee is unlikely to leave employment or move to private sector employment. Keeping the civil servant employed for more years costs the government more in wages. Given that civil service wage scales are typically tightly linked to seniority, this is an increase in the number of the most highly paid employees, resulting in a substantial increase in the wage bill. On the pension side, there is a reduction in the number of pension beneficiaries, but note that because of the increased seniority and the higher wages, the pension when it does get paid will be higher. Since civil service pensions are more often linked to last salary then private sector pensions, the higher wages before retirement will result in much higher pensions (higher both because of the greater years of service and because of the higher wage on which they are based) paid for fewer years. From the perspective of the pension fund, if it exists, this is likely to be deficit-reducing. But from the overall Government’s perspective, it is not 26 clear. The Government, as a whole will have to pay the worker a high wage and employer contributions on that wage instead of paying a pension that might have been only 50 percent of the previous lower wage. Depending on a number of factors, including the worker’s productivity, the steepness of the seniority based wage profile, the level of employer contributions, the benefit generosity in the pension system, and the tax treatment of pensions and wages, the Government, as a whole, may be better off letting the worker retire at a younger age, as long as it is not ridiculously young. Reforms that result in lower benefits affect both private sector pension systems and civil service pension systems positively. Reforms like reducing the indexation of the pension system to inflation, increasing the averaging period of the pensionable wage base, and lowering the accrual rate will all have positive impacts on the private sector system, civil service pension systems, the pension fund finances, and the overall government fiscal balance. Pension funds may urge that one pension system and a new, reformed pension system opened. As noted above, no matter how generous or actuarially unbalanced a pension scheme may be, in the beginning it typically generates surpluses. As the pension system matures, these surpluses shrink and can turn into deficits, particularly if the scheme is too generous. Since pension fund administrators do not like to oversee a deficit-laden scheme, they often propose closing the old scheme, passing the liabilities on to the government and opening a new, reformed scheme with more fiscally sustainable parameters. They argue that they should not contaminate the new, more actuarially fair scheme with the liabilities of the old scheme. While this might make sense from the pension fund’s perspective, it makes little sense for the government, which would inherit the liabilities of the old scheme while the contributions from new contributors generate surpluses for the pension fund. Even if the new pension scheme starts out actuarially balanced, the same political pressures that made the old system more generous over time are likely to affect the new scheme, and once it faces maturity, it too is likely to face deficits. In some countries, this game has been played repeatedly, with the liabilities of as many as four successive pension schemes passed to the 27 government as new pension schemes are begun. Unless contributions from current workers are invested in individual, portable accounts, they should always be used to pay benefits to the retired, disabled, and survivors before general government resources are tapped. Reform can still take place, with successive cohorts of civil servants facing different accrual rates or different averaging periods for their pensionable wage base. Reform in most cases does not require that the current system be closed and a new one opened. The only reform that might require closing the current pension system is a move from a defined benefit system to a fully funded, defined contribution system. 9 Most systems operate on the basis of a defined benefit system, in which contributions from current workers are taken in by the government or the pension fund in return for benefits defined as a proportion of salary per year of service. At the end of the day, if the pension fund does not have sufficient funding to pay the promise, the government provides the resources. In a fully funded, defined contribution system, by contrast, the individual’s contributions are invested and are owned by the individual, rather than the government or pension fund. Pensions are paid from the amount of money in an individual’s account at retirement, from contributions and the interest earned on those contributions. The government typically has no liability to the pensioner, as the pensioner receives only what is in the account. If a pension system transitions completely from a defined benefit system to a fully funded, defined contribution system, the liabilities from the old defined benefit system are indeed passed on to the government. The government takes those liabilities now in return for avoiding future contingent liabilities for workers in the new system. In the case of transition to a mixed system, the Government takes on part of the liabilities of the old system while some continue to be financed by the portion of the contribution going to the defined benefit. 9 Examples include Kenya and Nigeria. 28 4. Conclusions Civil service pension reform clearly needs to be on the agenda in Sub-Saharan African countries, as its costs are beginning to squeeze out other budget expenditures. Ideally countries want to think about integrating their civil service pension systems with their national pension systems to promote a unified and flexible labor market. However, in many cases the national systems are not developed or comprehensive enough to incorporate the civil service. In these cases, the civil service pension reform may need to be undertaken as a self-standing reform. When undertaking civil service reform separately from the national system, practitioners need to focus on the overall impact on the government, not on the finances of the pension fund. In national system reforms, almost anything that improves the finances of the pension fund improves the finances of the government, since the government is responsible for financing deficits of the pension fund. In integrated systems, even though civil servants are included it is still true that whatever is good for the pension fund finances improves government finances, since there are typically more private sector workers than civil servants. However, in civil servant systems, improving a pension fund’s finances may have net negative impacts on overall government finances. For example, raising contribution rates will impose costs on the government as employer; and raising the retirement age can force a government to continue paying high salaries to less productive workers  and to pay these workers even higher benefits when they retire, which can also be negative for government budgets. Civil service pension funds also may conflict with government priorities as they attempt to maintain actuarial balance. One typical method for improving actuarial balance is to declare a pension system closed, pass its liabilities to the government, and start fresh with new contributions and no old liabilities. While this might improve the finances of the pension fund, it makes government finances much worse and should be avoided at all costs. In fact, it calls into question the rationale for having a separate civil service pension fund. It would be better 29 for the government to have contributions flow into the Treasury with pensions paid directly from the Treasury. Reforms to civil service systems are also tricky because they require civil servants to enact reforms that might not be in their own personal interest. Given that most civil service pension schemes are relatively generous, reforms will likely require some reduction in benefits. From a political economy standpoint, civil servants are being asked to propose, lobby for, and implement reforms that will reduce the generosity of their own benefits. Even the members of parliament who have to vote for the reforms are often covered by the schemes. Too often, the temptation is to apply reform parameters only to new entrants, with the result that savings accrued under these reforms will be realized in the far distant future, long after problems become acute. Despite these challenges, civil service pension reform is too important to be postponed. 30 References Dorfman, M. 2015. "Pension Patterns in Sub-Saharan Africa," Social Protection & Labor Discussion Paper No. 1503, World Bank, Washington, DC. Filmer, D., and L. Fox. 2014. Youth Employment in Sub-Saharan Africa. A copublication of the Agence Francaise de Developpement and the World Bank. Washington, DC: World Bank. IMF (International Monetary Fund). 2015. "World Economic Outlook Database," October, IMF, Washington, DC. Majoke, Z, and R. Palacios. 2014. “Assessing Mortality Differentials: The Case of Pension Fund Members in Ghana”, August (preliminary). OECD (Organisation for Economic Co-operation and Development). 2015. Pensions at a Glance 2015: OECD and G20 indicatord. OECD: Paris. Palacios, R. Forthcoming. "Ethiopia’s Pension System: Background, Current Status and Prospects for Future Expansion," working paper, World Bank, Washington, DC. PenCom (Nigeria National Pension Commission). 2016. Retrieved from http://www.pencom.gov.ng/index.php. Sluchynskyy, O. 2015. "Defining, Measuring, and Benchmarking Administrative Expenditures of Mandatory Social Security Programs," Social Protection & Labor Discussion Paper No. 1501, World Bank, Washington, DC. United Nations. 2015. World Population Prospects. The 2015 Division. Key Findings & Advance Tables. Department of Economic and Social Affairs, Population Division. New York: United Nations. Palacios, R., and E. Whitehouse. 2006. "Civil-service Pension Schemes Around the World," Social Protection and Discussion Paper No. 0602, World Bank, Washington, DC. 31 WillisTowersWatson. 2014. "Nigeria: Changes in 2014 Pension Reform Act Increase Need for Preparation and Communication," Retrieved from WillisTowersWatson Global News Briefs, August 27. World Bank. 2014. World Bank Workshop on Coherent Pension Policy in Africa, Accra, Ghana, December. World Bank. 2016. World Development Indicators, Washington, DC: World Bank. Zviniene, A., and M. Guven. 2015. "Cabo Verde: Long Term Projections of the Pension System," Working paper, World Bank, Washington, DC. 32 Annex A. Civil Service Pension Scheme Contribution Rates (% Wage) 40% 35% 30% 25% 20% 15% 10% 5% 0% Central African Republic Mauritania Civil Service Average Ethiopia Madagascar Cote d'Ivoire Zambia Burkina Faso Kenya Senegal National Average Lestho Sao Tome and Principe Botswana Benin Seychelles Rwanda Ghana Tanzania Guinea Bissau Namibia Niger South Africa Cabo Verde South Sudan Mali Togo Chad Swaziland Eritrea Employer Employee Source: Administrative Data for Respective Countries. B. Financing Mechanisms of Civil Service Pension Schemes Country Civil Service Scheme Contribution/Financing Mechanism Employer Employee Total Uganda budget Zimbabwe 5%-7% 0 5%-7% Mauritius budget 6% Angola Budget 7% Burundi Budget Cameroon Budget 10% Congo, Dem. Rep. Budget Gambia, The Budget Malawi Budget Mozambique Budget 7% Liberia Budget Source: Administrative Data for Respective Countries. 33 C. Evolution of Civil Service Pension Expenditures in Kenya 30.00 25.00 20.00 KSHS, Billion 15.00 10.00 5.00 - Source: World Bank Workshop on Coherent Pension Policy in Africa, Accra Ghana, December 2014. 34 Social Protection & Labor Discussion Paper Series Titles 2014-2016 No. Title 1613 Issues for Civil Service Pension Reform in Sub-Saharan Africa by Anita M. Schwarz and Miglena Abels, November 2016 1612 How to Target Households in Adaptive Social Protection Systems? Relative Efficiency of Proxy Means Test and Household Economy Analysis in Niger by Pascale Schnitzer, October 2016 1611 Pensions for Public-Sector Employees: Lessons from OECD Countries’ Experience by Edward Whitehouse, October 2016 1610 Pension Systems in Sub-Saharan Africa: Brief Review of Design Parameters and Key Performance Indicators by Miglena Abels and Melis U. Guven, October 2016 1609 Household Enterprises in Fragile and Conflict-Affected States: Results from a Qualitative Toolkit Piloted in Liberia, Volume 2 – Annexes by Emily Weedon and Gwendolyn Heaner, August 2016 1608 Household Enterprises in Fragile and Conflict-Affected States: Results from a Qualitative Toolkit Piloted in Liberia, Volume 1 – Report by Emily Weedon and Gwendolyn Heaner, August 2016 1607 Benefits and Costs of Social Pensions in Sub-Saharan Africa by Melis U. Guven and Phillippe G. Leite, June 2016 1606 Assessing Benefit Portability for International Migrant Workers: A Review of the Germany- Turkey Bilateral Social Security Agreement by Robert Holzmann, Michael Fuchs, Seçil Paçacı Elitok and Pamela Dale, May 2016 1605 Do Bilateral Social Security Agreements Deliver on the Portability of Pensions and Health Care Benefits? A Summary Policy Paper on Four Migration Corridors Between EU and Non-EU Member States by Robert Holzmann, May 2016 1604 Assessing Benefit Portability for International Migrant Workers: A Review of the France-Morocco Bilateral Social Security Agreement by Robert Holzmann, Florence Legro and Pamela Dale, May 2016 1603 Assessing Benefit Portability for International Migrant Workers: A Review of the Belgium-Morocco Bilateral Social Security Agreement by Robert Holzmann, Jacques Wels and Pamela Dale, May 2016 1602 Assessing Benefit Portability for International Migrant Workers: A Review of the Austria-Turkey Bilateral Social Security Agreement by Robert Holzmann, Michael Fuchs, Seçil Paçaci Elitok and Pamela Dale, May 2016 1601 The Greek Pension Reform Strategy 2010-2016 by Georgios Symeonidis, July 2016 1507 Integrating Disaster Response and Climate Resilience in Social Protection Programs in the Pacific Island Countries by Cecilia Costella and Oleksiy Ivaschenko, September 2015 1506 Effectiveness of Targeting Mechanisms Utilized in Social Protection Programs in Bolivia by Ignacio Apella and Gastón Blanco, September 2015 1505 Kyrgyz Republic: Social Sectors at a Glance by João Pedro Azevedo, Paula Calvo, Minh Nguyen and Josefina Posadas, August 2015 1504 Entering the City: Emerging Evidence and Practices with Safety Nets in Urban Areas by Ugo Gentilini, July 2015 1503 Pension Patterns in Sub-Saharan Africa by Mark Dorfman, July 2015 1502 Social Protection in Fragile and Conflict-Affected Countries: Trends and Challenges by Mirey Ovadiya, Adea Kryeziu, Syeda Masood and Eric Zapatero, April 2015 1501 Defining, Measuring, and Benchmarking Administrative Expenditures of Mandatory Social Security Programs by Oleksiy Sluchynsky, February 2015 1425 Old-Age Financial Protection in Malaysia: Challenges and Options by Robert Holzmann, November 2014 1424 Profiling the Unemployed: A Review of OECD Experiences and Implications for Emerging Economies by Artan Loxha and Matteo Morgandi, August 2014 1423 Any Guarantees? An Analysis of China’s Rural Minimum Living Standard Guarantee Program by Jennifer Golan, Terry Sicular and Nithin Umapathi, August 2014 1422 World Bank Support for Social Safety Nets 2007-2013: A Review of Financing, Knowledge Services and Results by Colin Andrews, Adea Kryeziu and Dahye Seo, June 2014 1421 STEP Skills Measurement Surveys: Innovative Tools for Assessing Skills by Gaëlle Pierre, Maria Laura Sanchez Puerta, Alexandria Valerio and Tania Rajadel, July 2014 1420 Our Daily Bread: What is the Evidence on Comparing Cash versus Food Transfers? by Ugo Gentilini, July 2014 1419 Rwanda: Social Safety Net Assessment by Alex Kamurase, Emily Wylde, Stephen Hitimana and Anka Kitunzi, July 2012 1418 Niger: Food Security and Safety Nets by Jenny C. Aker, Carlo del Ninno, Paul A. Dorosh, Menno Mulder-Sibanda and Setareh Razmara, February 2009 1417 Benin: Les Filets Sociaux au Bénin Outil de Réduction de la Pauvreté par Andrea Borgarello et Damien Mededji, Mai 2011 1416 Madagascar Three Years into the Crisis: An Assessment of Vulnerability and Social Policies and Prospects for the Future by Philippe Auffret, May 2012 1415 Sudan Social Safety Net Assessment by Annika Kjellgren, Christina Jones-Pauly, Hadyiat El-Tayeb Alyn, Endashaw Tadesse and Andrea Vermehren, May 2014 1414 Tanzania Poverty, Growth, and Public Transfers: Options for a National Productive Safety Net Program by W. James Smith, September 2011 1413 Zambia: Using Social Safety Nets to Accelerate Poverty Reduction and Share Prosperity by Cornelia Tesliuc, W. James Smith and Musonda Rosemary Sunkutu, March 2013 1412 Mali Social Safety Nets by Cécile Cherrier, Carlo del Ninno and Setareh Razmara, January 2011 1411 Swaziland: Using Public Transfers to Reduce Extreme Poverty by Lorraine Blank, Emma Mistiaen and Jeanine Braithwaite, November 2012 1410 Togo: Towards a National Social Protection Policy and Strategy by Julie van Domelen, June 2012 1409 Lesotho: A Safety Net to End Extreme Poverty by W. James Smith, Emma Mistiaen, Melis Guven and Morabo Morojele, June 2013 1408 Mozambique Social Protection Assessment: Review of Social Assistance Programs and Social Protection Expenditures by Jose Silveiro Marques, October 2012 1407 Liberia: A Diagnostic of Social Protection by Andrea Borgarello, Laura Figazzolo and Emily Weedon, December 2011 1406 Sierra Leone Social Protection Assessment by José Silvério Marques, John Van Dyck, Suleiman Namara, Rita Costa and Sybil Bailor, June 2013 1405 Botswana Social Protection by Cornelia Tesliuc, José Silvério Marques, Lillian Mookodi, Jeanine Braithwaite, Siddarth Sharma and Dolly Ntseane, December 2013 1404 Cameroon Social Safety Nets by Carlo del Ninno and Kaleb Tamiru, June 2012 1403 Burkina Faso Social Safety Nets by Cécile Cherrier, Carlo del Ninno and Setareh Razmara, January 2011 1402 Social Insurance Reform in Jordan: Awareness and Perceptions of Employment Opportunities for Women by Stefanie Brodmann, Irene Jillson and Nahla Hassan, June 2014 1401 Social Assistance and Labor Market Programs in Latin America: Methodology and Key Findings from the Social Protection Database by Paula Cerutti, Anna Fruttero, Margaret Grosh, Silvana Kostenbaum, Maria Laura Oliveri, Claudia Rodriguez-Alas, Victoria Strokova, June 2014 To view Social Protection & Labor Discussion papers published prior to 2013, please visit www.worldbank.org/spl. Abstract The paper summarizes the main factors behind the projected increase civil service pension costs in Sub-Saharan Africa (SSA). It discusses the benefits and potential downsides to unifying civil service and national pension systems, drawing on regional and international best practices and experience. The paper pays special consideration to the differences between civil and national pension reform, emphasizing the unique challenges in civil service pension reform posed by the fact that the government is the employer, the administrator of the pension fund, and the guarantor of last resort of the pension system. Findings in the report strongly suggest that civil service pension reform needs to be on the agenda in SSA countries, as its costs are beginning to crowd out other budget expenditures. Among other conclusions and recommendations, the report also urges practitioners to focus on the overall impact on government finances and not on the finances of the pension fund when undertaking civil service pension reform separately from the national system. The paper is intended to serve as a resource in civil service pension reform efforts in the region. About this series Social Protection & Labor Discussion Papers are published to communicate the results of The World Bank’s work to the development community with the least possible delay. This paper therefore has not been prepared in accordance with the procedures appropriate for formally edited texts. The findings, interpretations, and conclusions expressed herein are those of the author(s), and do not necessarily reflect the views of the International Bank for Reconstruction and Development/The World Bank and its affiliated organizations, or those of the Executive Directors of The World Bank or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgement on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. For more information, please contact the Social Protection Advisory Service, The World Bank, 1818 H Street, N.W., Room G7-803, Washington, DC 20433 USA. Telephone: (202) 458-5267, Fax: (202) 614-0471, E-mail: socialprotection@worldbank.org or visit us on-line at www.worldbank.org/spl. © 2016 International Bank for Reconstruction and Development / The World Bank